JP2007509649A6 - Local path automatic planning method and apparatus for virtual colonoscopy - Google Patents

Abstract

  The present invention relates to a local path automated planning method for virtual endoscopy, deriving a colon data set obtained by a colon examination protocol for use in subsequent steps, Determining a sub-volume around an endoscope position; executing a region growing within the lumen starting from the current endoscope position; and a plurality of cube faces circumscribing the sub-volume Calculating and clustering the intersection of the region with the region, calculating an approximate centerline path from the current endoscope position to the center of each cluster formed in the preceding step, and the centerline path Comparing each to a current path indicated by an endoscope, and for each of the centerline paths Assigning a score based on the comparison and includes the steps of selecting a centerline path based on the scores.

Description

  This application was filed with the US Patent and Trademark Office on 30 January 2003 under the title “Automatic Local Path Planning for Virtual Colonoscopy” and inventor “Bernhard Geiger”, US Provisional Patent Application No. 60/443734. Claims priority based on the specification. This application is also filed with the United States Patent and Trademark Office on May 14, 2003 under the title “Fast Centerline Extraction” and inventor “Bernhard Geiger / Jean Daniel Boissonnat”, US Provisional Patent Application No. 60/470579. Claims priority based on the specification.

  The present invention basically relates to computer vision imaging systems, and in particular, automatic planning of local paths as used in virtual endoscopy and virtual colonoscopy. The present invention relates to a method and apparatus (system).

BACKGROUND ART Virtual colonoscopy (VC) is a diagnosis based on computer simulation of a standard minimally invasive endoscopic procedure using a patient-specific anatomical three-dimensional (3D) data set. It is about the method. Examples of current endoscopic procedures include bronchoscopy, sinus endoscopy, upper gastrointestinal endoscopy, colonoscopy, cystoscopy, cardiac endoscopy And urethoscopy. VC visualization of patient-specific anatomical structures obtained non-invasively avoids the following risks: perforation, infection, bleeding, etc., and is important prior to actual endoscopy Information is provided to the endoscope operator. Such knowledge minimizes procedural difficulties, reduces patient morbidity, increases practicability, and improves a better understanding of treatment results.

  In virtual colonoscopy, 3D images are created from, for example, volume rendering from two-dimensional (2D) computer tomography (CT) data or nuclear magnetic resonance (MR) data.

  These 3D images are created to simulate images coming from an actual endoscope, such as a fiber optic endoscope. This means that the viewpoint of the virtual endoscope must capture the lumen of an organ or other human tissue from the inside with a wide viewing angle (typically about 100 degrees arc). This viewpoint must be moved along the interior of the lumen, which means that a stereoscopic (3D) translational movement and a stereoscopic (3D) rotation must be provided. Controlling these parameters interactively is a challenge.

  A commonly used technique for navigating the virtual endoscope viewpoint is to calculate a “flight path” in advance and automatically move the virtual colonoscope viewpoint along this path. However, this technique requires a segmentation and trajectory calculation step with potential for failure and time.

The contents of the previous application No. 10/322, 326 filed on December 18, 2002 under the title "Automatic Navigation for Virtual ENdoscopy" are not here compatible with the present application. It is taken up as a reference document that is not. Disclosed herein is an internal colon dataset guidance system and method that uses the longest line of sight for advancement. This earlier application shows a system and method for automatic guidance of the endoscopic viewpoint in virtual colonoscopy. This system automatically determines the direction and orientation of the virtual endoscope. The user therefore only needs to control only one parameter, ie forward or backward speed, and this method is directly interactive inside the organ without any pre-processing (ie segmentation or path creation). It enables guidance. The method of the application includes the following steps. That is,
a) determining an initial viewpoint of the virtual endoscope, the initial viewpoint having a first center point and a first direction; b) determining a longest line of sight from the initial viewpoint with respect to the lumen; The longest line of sight has a first longest line of sight direction; c) determining a second direction between the first direction of the initial viewpoint and the first longest line of sight direction; d) turning the viewpoint to the second direction; Moving the initial viewpoint in the first direction by a predetermined first interval; e) calculating a second center point of the viewpoint; f) moving the viewpoint to the second center point; and from steps b) to f ) Until the viewpoint reaches the intended target.

SUMMARY OF THE INVENTION In operation of the system disclosed in the aforementioned earlier application (Attorney Docket No. 2001P24461US), virtual colonoscopy, under certain circumstances, bends at a right angle and stagnates in deep areas. The possibility of folds due to swelling is observed.

  According to one aspect of the present invention, a method is provided for navigating a virtual colonoscopic viewpoint within a lumen of a tissue. According to an aspect of the invention, a method for automatic path planning includes the following steps. That is, a step of determining a sub-volume around the current endoscope position, a step of executing a region growing inside the lumen starting from the endoscope position, and six surfaces of a cube circumscribing the region of interest Calculating and clustering the intersection of the growth area with each other, calculating a rough center line from the endoscope position to the center of each cluster, and comparing each path with the current path of the endoscope And detecting the best score.

According to another aspect of the present invention, an automatic local path planning method for virtual endoscopy includes the following steps. That is,
Deriving a colon data set obtained by a colon examination protocol for use in subsequent steps, determining a sub-volume around the current endoscopic position in the lumen, and the current endoscopic position Starting with a region that grows within the lumen, calculating and clustering intersections of the cube with a plurality of faces circumscribing the subvolume, and the current endoscope position Calculating a rough centerline path from each to the center of each cluster formed in the preceding step, comparing each of the centerline paths to a current path indicated by an endoscope, and the centerline Assigning a score based on the comparison to each of the paths, and a centerline based on the score And a step of selecting a scan.

  According to another aspect of the invention, deriving a colon dataset obtained by the colon examination protocol includes deriving the dataset by computer tomography.

  According to another aspect of the present invention, deriving a colon data set obtained by the colon examination protocol includes deriving the data set by nuclear magnetic resonance.

  Yet another aspect of the invention includes the step of determining a cube that leads a predetermined number of voxels centered around the current endoscope.

  According to another aspect of the present invention, calculating the approximate centerline path includes calculating an initial path and centering and smoothing the initial path.

  According to another advantageous embodiment of the invention, the step of centering and smoothing the initial path includes using Gaussian smoothing.

  According to yet another embodiment of the present invention, the step of centering and smoothing the initial path, the step of setting a sphere at the apex position, and the size of the sphere until a collision with a lumen wall occurs. Enlarging, calculating a transition force from the collision, supplying the transition force until the sphere no longer collides, and determining the size of the sphere to further collide with the lumen wall Expanding until the first time occurs again, calculating a further transition force, supplying the additional transition force until the sphere no longer collides, and the preceding three steps: Until the final position is reached at which no further growth is possible, and the final position is It includes a step of displaying a vertex position.

According to yet another embodiment of the present invention, a local path automatic planning method for virtual endoscopy includes the following steps. That is,
Deriving a colon data set obtained by a colon examination protocol for use in subsequent steps, determining a sub-volume around the current endoscopic position in the lumen, and the current endoscopic position Starting with a region that grows within the lumen, calculating and clustering intersections of the cube with a plurality of faces circumscribing the subvolume, and the current endoscope position Calculating an approximate centerline path from each to the center of each cluster formed in the preceding step; in this case, further comprising calculating an initial path, and centering and smoothing the initial path And comparing each of the centerline paths with the current path shown by the endoscope A step that includes the steps of assigning a score based on the comparison for each of the centerline path includes selecting a center line path based on the score is.

According to another aspect of the present invention, the following method steps are included in a local path automatic planning method for virtual endoscopy. That is,
Determining a subvolume around a current endoscope position in the lumen; executing a region starting from the current endoscope position and growing inside the lumen; and circumscribing the subvolume Calculating and clustering intersections between a plurality of surfaces of the cube to be processed and the region, and calculating an approximate centerline path from the current endoscope position to the center of each cluster formed in the preceding step Comparing each of the centerline paths with a current path indicated by an endoscope, and selecting an optimum centerline path based on the comparison.

According to another aspect of the present invention, an automatic local path planning system for virtual endoscopy is configured to include the following means. That is,
Means for determining a subvolume around a current endoscope position in the lumen; means for executing a region starting from the current endoscope position and growing inside the lumen; and circumscribing the subvolume Means for calculating and clustering intersections between a plurality of surfaces of a cube to be processed and the region, and means for calculating an approximate centerline path from the current endoscope position to the center of each cluster formed in a preceding step And means for comparing each of the centerline paths with the current path indicated by the endoscope, and means for selecting the optimum centerline path based on the comparison.

According to another aspect of the present invention, an automatic local path planning system for virtual endoscopy is configured to include the following means. That is,
Means for deriving a colon data set obtained by a colon examination protocol for use in subsequent steps; means for determining a sub-volume around a current endoscopic position in the lumen; and the current endoscopic position Means for executing a region growing within the lumen starting from the step, means for calculating and clustering intersections of the plurality of faces circumscribing the subvolume and the region, and the current endoscope position Means for calculating a rough centerline path from each to the center of each cluster formed in the preceding step, means for comparing each of the centerline paths with a current path indicated by an endoscope, and the centerline Means for assigning a score based on the comparison to each of the paths; and means for selecting a centerline path based on the score. Is constructed sea urchin.

  According to yet another aspect, means for centering and smoothing the initial path, means for setting a sphere at the apex position, means for expanding the size of the sphere until a collision with a lumen wall occurs, Means for calculating the transition force from the collision; means for supplying the transition force until the sphere no longer falls into collision; and the size of the sphere until a further collision with the lumen wall occurs again. Means for enlarging, means for calculating further transition forces, means for supplying said further transition forces until the sphere no longer collides, and the preceding three steps: Means for repeating until reaching a final position where growth cannot be achieved and means for displaying the final position as the final vertex position are included.

  According to yet another aspect, a local path automatic planning method for virtual endoscopy includes the following steps. A step of determining a subvolume around a current endoscope position in a lumen; a step of executing a region starting from the current endoscope position and growing inside the lumen; and the subvolume Calculating and clustering the intersection of a plurality of cube faces circumscribing to the region, and calculating the approximate centerline path from the current endoscope position to the center of each cluster formed in the preceding step Comparing each of the centerline paths with a current path indicated by an endoscope, and selecting an optimal centerline path based on the comparison.

BRIEF DESCRIPTION OF THE DRAWINGS For a more complete understanding of the various aspects of the present invention, embodiments of the invention are described in detail below with reference to the drawings. In this case, FIG. 1 is a diagram schematically showing a typical situation in which the progress of the endoscope is stagnant at a bending point in order to support a complete understanding of the present invention.
2 and 3 are diagrams showing a voxel path for centerline extraction that can be used in an embodiment of the present invention.
FIG. 4 shows a centering step that can be used in an embodiment of the present invention.

Embodiments It should be understood that the method and system of the present invention is implemented using a programmable digital computer, and that the operations described therein are also related to the computer implementation. In the imaging context, terms such as “air”, “lumen / lumen” are typically meant to refer to the corresponding imaging of these features.

  As described above, under certain circumstances, stagnation at the bends may occur in the operation of conventional systems for virtual colon examination, or deep bulges may occur in virtual colon examination. FIG. 1 shows a typical situation in which the endoscope is stagnant at a bending point.

  According to the embodiment of the present invention, the growth region from the endoscope position intersects the cube in the three clusters C0, C1, and C2. The system calculates a new centerline of the endoscope towards the center of each cluster and compares each path with the path of the endoscope. Here, a score that reflects the path difference from the input path is calculated. The highest scoring path is the path that continues most (extends) towards the colon. If there is only one cluster, the system determines that there is a dead end and a dead end has been reached.

  According to the embodiment based on the automatic path planning method according to the present invention, first, the step of defining the sub-volume space around the current endoscope position is included. Next, a step of performing region growth inside the lumen is performed, starting from the current endoscope position. This is followed by the steps of computing and clustering the intersection of the growth region and the six faces of the cube that circumscribe the region of interest. A rough centerline from the endoscope position for the center of each cluster is calculated. Each pass is compared to the current pass of the endoscope and the highest score is found.

The approximate centerline calculation step can be performed in different ways. Most centerline algorithms use decimation, morphological operations, distance transformation, minimum cost path, “Dijkstra's” algorithm, etc. Examples of these backgrounds include the following documents.
Zhou et al., “Three-Dimensional Skeleton and Centerline Generation Based on an Approximate Minimum Distance Field,” “The Visual Computer, 14: 303-314 (1998); R. Truyen,”, T. Deschamps, LDCohen, “Clinical evaluation of an automatic path tracker for virtual colonoscopy, Medical Image Computing and Computer-Assisted Intervention (MICCAI), Utrecht, Netherlands, October 2001, Chen et al., “A Fast Algorithm to Generate Centerline for Virtual Colonoscopy, SPIE Conference, Feb 12-18, 2000 ”, Richard Robb,“ Virtual (Compted) Endoscopy: Development and Evaluation Using the Visible Human Datasets, ”Oct. 7-8, 1996, www.mayo, edu; USPatent 6,514,082, entitiled“ System and method for performing a three-dimensional examination with collapse correction ”

  Other methods of performing colon segmentation include performing start and end point calculations and initial path calculations as described in the aforementioned provisional application 60/470579. This continues with centering and smoothing of the path, which begins with a colon data set obtained by use of typical colonoscopy protocols including intestinal preparations and air infusion. This data set is segmented by the use of thresholds for air and analysis of connected elements. In this case, elements that do not belong to the colon are truncated either automatically or manually. Starting with the first voxel belonging to the colon, voxel interval labeling is performed, typically with continuous numbering. For example, label 0 is given to the first voxel, label 1 is given to the next adjacent label, and label 2 is given to the next adjacent label. The search is performed on the highest label that is converted to the origin p0. From this starting point p0, a new interval label map is created by repeating the preceding steps of interval labeling and acquiring other voxels with the highest number. This is the end point p1.

  From the end point p1, the interval label is used to obtain the path of the connected voxel that ends at the start point p0. This is done by searching for voxels with smaller labels among the neighboring labels of p2, which continues until p0 is reached. 2a and 3a represent the initial voxel path. The resulting path is basically jagged and is therefore smoothed using well-known techniques such as Gaussian smoothing. Some vertices (or nodes) are replaced by a weighted average of n adjacent labels, and the process is repeated for a number of iterations. Several new vertex locations are tested for each collision with the intestinal wall by verifying that the new coordinates are still present in the segmented colon. In a collision event, the vertex leaves at the last position where there is no collision. In a sense, this process can be compared to a flexible, unweighted string path that extends through the colon. An initial smoothing step is shown in FIGS. 2b and 3b. Figures 2c and 3c represent the final centering.

  According to the embodiment described above, this smoothing pass is centered using a sphere with an enlarged size. FIG. 4 shows how a small sphere is centered at the apex along the path. A plurality of vertices on the sphere are checked for each collision with the intestinal wall. When a plurality of vertices are in a collision state, a force for moving the sphere away from the intestinal wall is determined based on the sphere normal. The sphere is forced to move on a plane perpendicular to the path. When the sphere is no longer in a collision state, the size of the sphere is increased and the collision calculation and shift are repeated again. This process is stopped when the sphere cannot be further shifted and / or increased in size without causing a collision. Here, the center of the sphere is accepted as a new position relative to the vertex. This process is repeated for the next vertex of the trajectory or path (see C in FIG. 1). For an overview of collision detection technology and migration force calculations, see “Real-Time-Collision” published by “Geiger, B” at Computer Graphics International 2000, June 19-23, 2000, in Geneva, Switzerland. Detection and Response for Complex Environments ”. However, collision detection and force calculation are performed directly based on voxels rather than polyhedral reconstruction. In other words, basically, the concept outlined in the aforementioned Geiger literature basically continues. Subsequent and after centering, the path is subjected to another Gaussian smoothing with collision control. This uses fewer iterations and smaller adjacent areas.

FIG. 4 shows a summarized procedure for the centering step. Ie:
a) The sphere is set at the apex position b) The sphere size is expanded to a collision with the intestinal wall. From the collision, the transition force is calculated c) This transition is supplied until the sphere no longer falls into collision. The size of the sphere is increased again, where d) the transition where collision with the intestinal wall occurs is calculated. After this transition, the sphere reaches a position where it cannot grow any further.
e) This position represents the final vertex position.

  According to the basic principle of the present invention, which does not use simple morphological operations, it is possible to reach a given situation more quickly.

According to an embodiment of the present invention, an automatic local path planning method for virtual colonoscopy includes steps as described below. That is,
First, automatic segmentation is performed, where the automatic segmentation is cut out into selected cube-sized blocks (eg, 128 × 128 × 128) and the voxels are centered at the position of the endoscope. When initiated by a voxel that includes an endoscope as a nucleus, a region that grows in the interior space is executed. All voxels that contain space and belong to the same lumen are labeled. This columnar space intersects the cube at various points. Those locations are clustered. In this case, the voxels facing in common are added to the same cluster, and the center of each cluster is calculated by averaging the coordinates of the voxels of each cluster. A center line from the center of each cluster to the position of the endoscope is calculated. Each centerline is compared to the endoscope path. Starting from the position of the endoscope, the absolute value of the difference between the points on the center line is added to the corresponding point in the endoscope path. The sum is divided by the length to give a score. A higher score is more separated from the centerline from the incoming path. The path with the highest score is the best candidate to follow. If there is only one path, it is the most promising one to insert the endoscope and the system detects the presence of a dead end.

  The present invention has been mainly described in the context of the most practical virtual colonoscopy, however, this does not mean that the present application is not suitable for other types of virtual endoscopy. I want you to understand. Although the present invention has been specifically described using exemplary embodiments, those skilled in the art will be able to understand the present invention as specified by the following claims in addition to these embodiments. It should also be understood that various modifications and variations are possible without departing from the spirit.

A diagram that schematically shows a typical situation where the progress of the endoscope is stagnant at the bending point. Figure representing a voxel path for centerline extraction that can be used in embodiments of the present invention. Figure representing a voxel path for centerline extraction that can be used in embodiments of the present invention. Diagram showing centering steps that can be used in an embodiment of the present invention

Claims (36)

In the local path automatic planning method for virtual endoscopy,
Deriving a colon data set obtained by a colon examination protocol for use in subsequent steps;
Determining a sub-volume around the current endoscopic position in the lumen;
Performing a region growing within the lumen starting from the current endoscope position;
Calculating and clustering intersections between a plurality of cubic faces circumscribing the subvolume and the region;
Calculating an approximate centerline path from the current endoscope position to the center of each cluster formed in the preceding step;
Comparing each of the centerline paths to a current path indicated by an endoscope;
Assigning a score based on the comparison to each of the centerline paths; and selecting a centerline path based on the scores.   The method of claim 1, wherein deriving a colon dataset obtained by the colon examination protocol includes deriving the dataset by computer tomography (CT).   The method of claim 1, wherein deriving a colon data set obtained by the colon examination protocol includes deriving the data set by nuclear magnetic resonance (MR).   The method of claim 1 including the step of determining a cube that leads a predetermined number of voxels centered around a current endoscope.   The method of claim 1, comprising forming a region that grows within a “space” within the lumen and labeling all voxels within the “space” within the lumen.   The method includes forming a cluster at each location where a space in the lumen intersects a corresponding plane of a cube, and including a voxel of a shared plane corresponding to the cluster in each cluster. the method of.   The method of claim 6, comprising calculating the center of each cluster by averaging the coordinates of the voxels of each cluster.   The method of claim 7 including calculating each centerline path from the center of each cluster relative to the current endoscope position. Assigning a score based on the comparison to each of the centerline paths;
Starting from the current endoscope position;
Forming a sum of absolute values of differences of a plurality of points of each centerline path with respect to corresponding points of the current path indicated by the endoscope;
The sum is divided by the length of each centerline path to form a commercial score (score);
The method of claim 1 including the step of selecting a path having the highest commercial score.   The method of claim 9, wherein there is only one centerline path and the step of displaying such a result as a dead end is included. Calculating the approximate centerline path;
Calculating an initial path;
The method of claim 1 including centering and smoothing the initial path. In the step of calculating the initial path,
12. The method of claim 11, comprising storing a plurality of voxels with increasing label numbers from the end point until the start point is reached.   The method of claim 11, wherein centering and smoothing the initial path includes using Gaussian smoothing. Centering and smoothing the initial path;
Setting the sphere at the apex position;
Expanding the size of the sphere until a collision with the lumen wall occurs;
Calculating a transition force from the collision;
Supplying the transition force until the sphere no longer collides;
Expanding the size of the sphere until a further collision with the lumen wall occurs again;
Calculating additional migration forces;
Supplying the further transition force until the sphere no longer collides;
Repeating the previous three steps until the sphere reaches a final position where it can no longer grow without collision.
And displaying the final position as a final vertex position. In the local path automatic planning method for virtual endoscopy,
Deriving a colon data set obtained by a colon examination protocol for use in subsequent steps;
Determining a sub-volume around the current endoscopic position in the lumen;
Performing a region growing within the lumen starting from the current endoscope position;
Calculating and clustering intersections between a plurality of cubic faces circumscribing the subvolume and the region;
Calculating an approximate centerline path from the current endoscope position to the center of each cluster formed in the preceding step; in this case, further calculating an initial path in the step; and Centering and smoothing the path,
Comparing each of the centerline paths to a current path indicated by an endoscope;
Assigning a score based on the comparison to each of the centerline paths; and selecting a centerline path based on the scores.   16. The method of claim 15, wherein deriving a colon data set obtained by the colon examination protocol includes deriving the data set by computer tomography (CT).   16. The method of claim 15, wherein deriving a colon data set obtained by the colon examination protocol includes deriving the data set by nuclear magnetic resonance (MR). In the local path automatic planning method for virtual endoscopy,
Determining a sub-volume around the current endoscopic position in the lumen;
Performing a region growing within the lumen starting from the current endoscope position;
Calculating and clustering intersections between a plurality of cubic faces circumscribing the subvolume and the region;
Calculating an approximate centerline path from the current endoscope position to the center of each cluster formed in the preceding step;
Comparing each of the centerline paths to a current path indicated by an endoscope;
Selecting an optimal centerline path based on the comparison. In the local path automatic planning system for virtual endoscopy,
Means for determining a subvolume around a current endoscopic position in the lumen;
Means for executing a region growing within the lumen starting from the current endoscopic position;
Means for computing and clustering intersections between a plurality of faces of a cube circumscribing the subvolume and the region;
Means for calculating a rough centerline path from the current endoscope position to the center of each cluster formed in the preceding step;
Means for comparing each of the centerline paths to the current path indicated by the endoscope;
And a means for selecting an optimum centerline path based on the comparison. In the local path automatic planning system for virtual endoscopy,
Means for deriving a colon data set obtained by a colon examination protocol for use in subsequent steps;
Means for determining a subvolume around a current endoscopic position in the lumen;
Means for executing a region growing within the lumen starting from the current endoscopic position;
Means for computing and clustering intersections between a plurality of faces of a cube circumscribing the subvolume and the region;
Means for calculating a rough centerline path from the current endoscope position to the center of each cluster formed in the preceding step;
Means for comparing each of the centerline paths to the current path indicated by the endoscope;
Means for assigning a score based on the comparison to each of the centerline paths;
And a means for selecting a centerline path based on the score.   21. The system of claim 20, wherein means for deriving a colon data set obtained by the colon examination protocol includes means for deriving the data set by computer tomography (CT).   21. The system of claim 20, wherein means for deriving a colon data set obtained by the colon examination protocol includes means for deriving the data set by nuclear magnetic resonance (MR).   21. The system of claim 20, including means for determining a cube that leads a predetermined number of voxels centered around a current endoscope.   21. The system of claim 20, including means for forming a region that grows within a "space" within the lumen and labeling all voxels within the "space" within the lumen.   Means for forming a cluster at each location where a space in the lumen intersects a corresponding plane of a cube and including a voxel of a shared plane corresponding to the cluster in each cluster. 20. The system according to 20.   26. The system of claim 25, comprising means for calculating the center of each cluster by averaging the coordinates of the voxels of each cluster.   27. The system of claim 26, including means for calculating each centerline path from the center of each cluster relative to the current endoscope position. Means for assigning a score based on the comparison to each of the centerline paths;
Means starting from the current endoscope position;
Means for forming a sum of absolute values of differences of a plurality of points of each centerline path with respect to corresponding points of the current path indicated by the endoscope;
Means by which the sum is divided by the length of each centerline path to form a commercial score (score);
21. The system of claim 20, including means for selecting a path having the highest commercial score.   29. The system of claim 28, wherein there is only one centerline path and means for displaying such a result as a dead end is included. Means for calculating the approximate centerline path include:
A means of calculating an initial path;
21. The system of claim 20, including means for centering and smoothing the initial path. In the means for calculating the initial path,
31. The system of claim 30, including means for continuously storing a plurality of voxels with increasing label numbers from the end point until the start point is reached.   32. The system of claim 30, wherein the means for centering and smoothing the initial path includes means for using Gaussian smoothing. Means for centering and smoothing the initial path,
Means for setting the sphere at the apex position;
Means for enlarging the size of the sphere until a collision with the lumen wall occurs;
Means for calculating a transition force from the collision;
Means for supplying the transition force until the sphere no longer collides;
Means for increasing the size of the sphere until a further collision with the lumen wall occurs again;
A means to calculate additional migration forces;
Means for supplying said further transition force until said sphere no longer collides;
Means for repeating the previous three steps until the sphere reaches a final position where no further growth can occur without collision and means for displaying the final position as the final vertex position. The system of claim 30. In the local path automatic planning system for virtual endoscopy,
Means for deriving a colon data set obtained by a colon examination protocol for use in subsequent steps;
Means for determining a subvolume around a current endoscopic position in the lumen;
Means for executing a region growing within the lumen starting from the current endoscopic position;
Means for computing and clustering intersections between a plurality of faces of a cube circumscribing the subvolume and the region;
Means for calculating an approximate centerline path from the current endoscope position to the center of each cluster formed in the preceding step; in this case, further means for calculating an initial path; Means for centering and smoothing the path,
Means for comparing each of the centerline paths to the current path indicated by the endoscope;
Means for assigning a score based on the comparison to each of the centerline paths;
And a means for selecting a centerline path based on the score. In the local path automatic planning system for virtual endoscopy,
Means for determining a subvolume around a current endoscopic position in the lumen;
Means for executing a region growing within the lumen starting from the current endoscopic position;
Means for computing and clustering intersections between a plurality of faces of a cube circumscribing the subvolume and the region;
Means for calculating a rough centerline path from the current endoscope position to the center of each cluster formed in the preceding step;
Means for comparing each of the centerline paths to the current path indicated by the endoscope;
And a means for selecting an optimum centerline path based on the comparison. In the local path automatic planning system for virtual endoscopy,
Means for determining a subvolume around a current endoscopic position in the lumen;
Means for executing a region growing within the lumen starting from the current endoscopic position;
Means for computing and clustering intersections between a plurality of faces of a cube circumscribing the subvolume and the region;
Means for calculating a rough centerline path from the current endoscope position to the center of each cluster formed in the preceding step;
Means for comparing each of the centerline paths to the current path indicated by the endoscope;
And a means for selecting an optimum centerline path based on the comparison.

Images